283 replies to this topic

[thievingmagpi]

Nightbird, on 30 June 2020 - 08:25 PM, said:

I've elucidated many times, with explanations posted as early as June 11th. As mentioned before, your group worked in a silo and the result shows.

Ah yes, the classic "it's their fault, not mine" lmao.

Over a week of almost constant discussion between dozens of people and you were unable to clearly, succinctly, accurately, effectively, whatever else-ly explain the suitability of your model.

Clearly, some kind of conspiracy against you.

[Nightbird]

You can lead a horse to water but you can't make him drink

[thievingmagpi]

Lazy people blame others- Michael Jordan

[Nightbird]

thievingmagpi, on 30 June 2020 - 08:37 PM, said:

Lazy people blame others- Michael Jordan

Agreed, but who is the lazy one here and who put in more effort? lol

[thievingmagpi]

And yet here we are without your system chosen, despite pretty much everyone agreeing it was well made. Must've been that weird gigantic tryhard conspiracy I keep hearing about again.

[Nightbird]

thievingmagpi, on 30 June 2020 - 08:42 PM, said:

And yet here we are without your system chosen, despite pretty much everyone agreeing it was well made. Must've been that weird gigantic tryhard conspiracy I keep hearing about again.

Like the first reply in this thread lying about how the system is outside PGI's scope? Yeah, it is all in my head.

[thievingmagpi]

Oh yes, that old conspiracy lying about you again, and not you being unable to clearly, succinctly, accurately,effectively, whatever else-ly explain the suitability of your model.

[Nightbird]

thievingmagpi, on 30 June 2020 - 08:54 PM, said:

Oh yes, that old conspiracy lying about you again, and not you being unable to clearly, succinctly, accurately,effectively, whatever else-ly explain the suitability of your model.

Conspiracy implies secrecy... what you're doing is out in the open mate

[Sjorpha]

I don't give a damn about whether Nightbird communicated perfectly or not, or whether the community discussion was handled well or not, or whether the disinformation spread that his system is out of scope was an intentional lie or just a mistake (I was fooled by this disinformation too so yes it definitely caused harm to the conversation).

I'm just going to assume everyone involved acted in good faith and wants better matchmaking.

What I do care about is the facts and how things are handled from here, and as far as I can see the facts are as follows:

- Nightbird has demonstrated that his system is not out of scope and presented his solutions.
- Nightbird has produced comparative simulations predicting how the suggested models will work over time, showing that the model currently chosen will probably not work well, and that his suggested solution will work better both in terms of long term sustainability and short term improvements to match quality.
- I can't see any convincing simulations presented to challenge these findings, probably because he's right. Correct me if I'm wrong here and point me to those simulations (I'm talking about simulations of the system over time with comparable quality to Nightbirds).
- Jay Z himself has admitted that these flaws are inherent to his suggestion, and suggested a corrective periodic measure to counteract this. But this is clearly a worse alternative to a system that actually works better, corrects itself and improves over time.

This should not be a popularity contest, so we should implement the system that objectively works best in terms of improving match quality over time. That's not a matter of opinion but of mathematics, and it should ultimately be done and decided by people who understand the math involved. If there was any rhyme or reason to this PGI should simply hire someone with the required professional skills to make this decision and implement it.

So it seems to me that unless someone can explain how Nightbirds simulations are flawed and provide convincing simulations that shows Nightbirds predictions wrong such that the chosen model actually outperforms Nightbirds, the only rational discourse is to retract the other suggestions and rally behind his suggested solution (or a new one with similar or better simulated results over time).

So that's my 2 cents on this.

Consider this: If you go ahead pushing for a solution that you yourselves know is objectively worse just because you dislike someones behaviour you are being very irrational about this, you don't want to be that person.

Edited by Sjorpha, 01 July 2020 - 06:14 AM.

[HolyGrail101]

We need a major overhaul to Matchscore if the new PSR is going to work. I'm a primary Light Mech pilot that enjoys playing Map and Mode. What is the point of Conquest or Base Capture? Without updating other structural components of the game the only mode for Tier 1's is now Brawl in the middle in the biggest mech possible.

Maybe make K/D less of a thing and make the actual Capping part in a Cap mode's more significant (Among other changes).

I watched 2 Firestarter's base cap win a match and each got less than 80 Matchscore for winning.

[caseysrevenge]

Those 2 Firestarters understood the mode they were playing , and youre right K/D ratio should have been abondoned years ago.
Team play should be great reward.

[Brauer]

HolyGrail101, on 01 July 2020 - 03:27 AM, said:

We need a major overhaul to Matchscore if the new PSR is going to work. I'm a primary Light Mech pilot that enjoys playing Map and Mode. What is the point of Conquest or Base Capture? Without updating other structural components of the game the only mode for Tier 1's is now Brawl in the middle in the biggest mech possible.

Maybe make K/D less of a thing and make the actual Capping part in a Cap mode's more significant (Among other changes).

I watched 2 Firestarter's base cap win a match and each got less than 80 Matchscore for winning.

It all depends on what you mean by playing map and mode. If what you mean is running around the map and sitting on cap points without fighting you are not actually (in most cases) contributing that much. The reason for that is that taking opposing mechs off the field by fighting makes them less able to damage your mechs and play the map and mode. Very often cappers end up in a 5v1 or worse at the end. So objectively running around sitting on cap points is often not the most effective tactic AND it does not take a huge amount of skill or effort.

If you watch some high level competitive matches you will see that strong teams play the map and mode effectively, however that is much more about controlling areas of the map (including a few caps) than it is about running around sitting on points. Unfortunately, you will not see the teamwork or cohesion necessary for those types of strategies in QP (ok sometimes groups can do somewhat analogous things, but that's rare).

[Grus]

Did they try to fix something by throwing it down a hill again?

[Thorn Hallis]

Just a band aid fix that means nothing outside of epeening. Game has not enough players anymore to make a new PSR system meaningful

[VonBruinwald]

Grus, on 01 July 2020 - 04:49 AM, said:

Did they try to fix something by throwing it down a hill again?

In a nutshell:

• PGI threw their hands up and said "Fine! You fix the 'wolfing matchmaker if you're so smart".
  
• So an elite-clique came up with a solution.
  
• Then Nightbird came up with a better solution.
  
• But he wasn't part of the in-crowd so the clique dragged his idea through the mud and presented theirs to PGI round the back. PGI implemented their idea.

And now:

• Nightbird has gone public with his idea.
  
• And the clique are attempting to drag it through the mud again, but the models are on his side.
  
• Which has left the clique salty, seeing as they pitched their idea as community driven.

[Xiphias]

Nightbird said:

It fits perfectly within the scope outlined by PGI, this is something I can easily prove.

thievingmagpi, on 30 June 2020 - 08:09 PM, said:

And yet you haven't.

Weird huh?

No one called your system wrong or bad. Far from it. Make some more threads about how no one listens to you why don't you?

Failing to properly elucidate the advantages, and more importantly, the method of implementation is your fault. No one else's. The legitimacy of underlying systems do not matter a single iota if you can't deliver their benefits properly.

Nightbird, on 28 June 2020 - 03:17 PM, said:

As I stated here, you can do a rolling WLR calculation by setting PSR=WLR, after every game returning RESULT (1 or 0) from the game server for win or loss, and

new_PSR = {99 * [old_PSR/(old_PSR+ 1)] + RESULT} / {99* [1 - old_PSR/(old_PSR+ 1)] - RESULT + 1}

The number 99 creates a rolling average of 100 games.

There is no need to poll long term stats, everything is done with the systems in place today.

If a WLR around 1 with lots of decimals doesn't suit people (calculators don't really care), then re-scale it by multipling by 2500.

WLR=1.1 >> PSR=2750.

Tier 5 = [0-1500] Tier 4 = [1500-2000] Tier 3 = [2000-2500] Tier 2 = [2500-3250] Tier 1 = [>3250 no cap]

I don't want to get into the larger argument between the two of you, but in the interest of fairness and finding the best solution to the problem hasn't NightBird demonstrated how his proposal can fit inside the system? As far as I'm aware it has the most robust math behind it as I haven't seen anyone else build a simulation and show the long term results of any of the other proposed systems to be any better (my very limited work playing with a simulation seemed to support Nighbird's approach).

Setting aside the way that NightBird has gone about having the discussion, what are your complaints about the actual system? The above should fit in the scope, it appears to have better results, and those results have been laid out (I feel at least) pretty clearly.

If this isn't a popularity contest and you don't have a vested interest in pushing any specific system, shouldn't you get behind this one and support it regardless of your personal feelings towards NightBird?

The only real complaint I've seen towards his system (other than how he's presented it and himself) is that the broader community won't accept a pure WLR based system. If that's the case, shouldn't we put our efforts towards convincing the community to accept the best solution, rather than pushing one that has known flaws?

[Sjorpha]

so much this ^

[thievingmagpi]

Sjorpha, on 01 July 2020 - 12:25 AM, said:

I don't give a damn about whether Nightbird communicated perfectly or not,

End of discussion then.

Any model whose suitability is not and/or cannot properly be demonstrated fails as a model.

Sjorpha, on 01 July 2020 - 12:25 AM, said:

I'm just going to assume everyone involved acted in good faith and wants better matchmaking.

And the system that was adopted was one which was appropriately communicated. It's really quite simple. Nightbird was given every opportunity over the course of however many days the request-for-help was open and was unable- and in many cases outright refused.

Sjorpha, on 01 July 2020 - 12:25 AM, said:

- Nightbird has demonstrated that his system is not out of scope and presented his solutions.

If that were the case, this wouldn't be an discussion and we'd be using his model. He acted like a butthurt child when people disagree with his model, and instead of revising, or even searching out a synergistic solution, he doubled down on his indignation. This was a debate that took place over the course of many days.

Sjorpha, on 01 July 2020 - 12:25 AM, said:

- Nightbird has produced comparative simulations predicting how the suggested models will work over time, showing that the model currently chosen will probably not work well, and that his suggested solution will work better both in terms of long term sustainability and short term improvements to match quality.

Nobody has ever as far as I know, questioned the efficacy of such a system. This is a moot point.

Sjorpha, on 01 July 2020 - 12:25 AM, said:

- Jay Z himself has admitted that these flaws are inherent to his suggestion, and suggested a corrective periodic measure to counteract this. But this is clearly a worse alternative to a system that actually works better, corrects itself and improves over time.

A system which cannot be adopted is a failure of system. Full stop. That there are flaws within any other is irrelevant, as there is no perfect solution.

Sjorpha, on 01 July 2020 - 12:25 AM, said:

Consider this: If you go ahead pushing for a solution that you yourselves know is objectively worse just because you dislike someones behaviour you are being very irrational about this, you don't want to be that person.

Understanding that a flawed, but generally improved system which works, is better than a good system that can't is important to understanding why the handful of choices were selected.


If you want to sit there and appeal to PGI for the umpteenth time to give us a "better" system, you're welcome to. These appeals have happened for years and they've consistently ignored them, to their own detriment- even when large scale community proposals with many hours of work have been offered. Assuming they're willing and/or capable of anything more than a basic, bandaid solution is just ignorance and completely counter productive.

The community was given a couple of days to develop and present a few ideas on new PSR systems which PGI could look at adopting in whole or in part. That was done. As far as I'm concerned, the last .xml file has been edited and the last spreadsheet validated until the servers are shut off.

Edited by thievingmagpi, 01 July 2020 - 11:10 AM.

[thievingmagpi]

VonBruinwald, on 01 July 2020 - 05:49 AM, said:

In a nutshell:

• PGI threw their hands up and said "Fine! You fix the 'wolfing matchmaker if you're so smart".
  
• So an elite-clique came up with a solution.
  
• Then Nightbird came up with a better solution.
  
• But he wasn't part of the in-crowd so the clique dragged his idea through the mud and presented theirs to PGI round the back. PGI implemented their idea.

And now:

• Nightbird has gone public with his idea.
  
• And the clique are attempting to drag it through the mud again, but the models are on his side.
  
• Which has left the clique salty, seeing as they pitched their idea as community driven.

I noticed you haven't yet addressed which of the dozens of participants in the discussion, people who submitted ideas, people who worked on mathematics etc you'd consider "elite" and which "community" platforms were apparently excluded from said discussion and final proposal.

Edited by thievingmagpi, 01 July 2020 - 10:53 AM.

[Nightbird]

I did try to work with the group in question:

Me: <posts math equations>
Group: that's just your opinion
Me: you're welcome to point out flaws in the math, and we can discuss it
Group: stop trying to push your solution, it's pointless

I have a screenshot of this. My one fault in communicating was the mistake of trying to seriously talk to a bunch of pertinacious kids.